The Use of Nonstationary Panel Time Series Data in the Analysis of Farmland Values

Time-series methods based on panel data are used to increase the power of conventional econometric tests of present value models commonly applied to studies of asset valuation. A second contribution is to allow for the presence of frictions in farmland markets by allowing for threshold nonlinearities in the empirical model. Using countylevel data on Iowa farmland prices and cash rental rates between 1987 and 2010, panel data unit root models that allow for switching-regimes provide evidence in favor of the present value models. In one regime, where a threshold in the spread between rents and values is yet to be reached, deviations are persistent and contribute to drive land prices away from their intrinsic values. Conversely, the other regime is characterized by strong mean reversion in which market valuation becomes closer to fundamental determinants of values.